proc.5: Note kernel version for /proc/PID/smaps VmFlags "wf" flag

[thirdparty/man-pages.git] / man5 / core.5

.\" Copyright (c) 2006, 2008 by Michael Kerrisk <mtk.manpages@gmail.com>
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.TH CORE 5 2019-10-10 "Linux" "Linux Programmer's Manual"
.SH NAME
core \- core dump file
.SH DESCRIPTION
The default action of certain signals is to cause a process to terminate
and produce a
.IR "core dump file" ,
a disk file containing an image of the process's memory at
the time of termination.
This image can be used in a debugger (e.g.,
.BR gdb (1))
to inspect the state of the program at the time that it terminated.
A list of the signals which cause a process to dump core can be found in
.BR signal (7).
.PP
A process can set its soft
.B RLIMIT_CORE
resource limit to place an upper limit on the size of the core dump file
that will be produced if it receives a "core dump" signal; see
.BR getrlimit (2)
for details.
.PP
There are various circumstances in which a core dump file is
not produced:
.IP * 3
The process does not have permission to write the core file.
(By default, the core file is called
.IR core
or
.IR core.pid ,
where
.I pid
is the ID of the process that dumped core,
and is created in the current working directory.
See below for details on naming.)
Writing the core file fails if the directory in which
it is to be created is nonwritable,
or if a file with the same name exists and
is not writable
or is not a regular file
(e.g., it is a directory or a symbolic link).
.IP *
A (writable, regular) file with the same name as would be used for the
core dump already exists, but there is more than one hard link to that
file.
.IP *
The filesystem where the core dump file would be created is full;
or has run out of inodes; or is mounted read-only;
or the user has reached their quota for the filesystem.
.IP *
The directory in which the core dump file is to be created does
not exist.
.IP *
The
.B RLIMIT_CORE
(core file size) or
.B RLIMIT_FSIZE
(file size) resource limits for the process are set to zero; see
.BR getrlimit (2)
and the documentation of the shell's
.I ulimit
command
.RI ( limit
in
.BR csh (1)).
.IP *
The binary being executed by the process does not have read
permission enabled.
.IP *
The process is executing a set-user-ID (set-group-ID) program
that is owned by a user (group) other than the real user (group)
ID of the process,
or the process is executing a program that has file capabilities (see
.BR capabilities (7)).
(However, see the description of the
.BR prctl (2)
.B PR_SET_DUMPABLE
operation, and the description of the
.I /proc/sys/fs/suid_dumpable
.\" FIXME . Perhaps relocate discussion of /proc/sys/fs/suid_dumpable
.\" and PR_SET_DUMPABLE to this page?
file in
.BR proc (5).)
.IP *
.I /proc/sys/kernel/core_pattern
is empty and
.I /proc/sys/kernel/core_uses_pid
contains the value 0.
(These files are described below.)
Note that if
.I /proc/sys/kernel/core_pattern
is empty and
.I /proc/sys/kernel/core_uses_pid
contains the value 1,
core dump files will have names of the form
.IR .pid ,
and such files are hidden unless one uses the
.BR ls (1)
.I \-a
option.
.IP *
(Since Linux 3.7)
.\" commit 046d662f481830e652ac34cd112249adde16452a
The kernel was configured without the
.BR CONFIG_COREDUMP
option.
.PP
In addition,
a core dump may exclude part of the address space of the process if the
.BR madvise (2)
.B MADV_DONTDUMP
flag was employed.
.PP
On systems that employ
.BR systemd (1)
as the
.I init
framework, core dumps may instead be placed in a location determined by
.BR systemd (1).
See below for further details.
.\"
.SS Naming of core dump files
By default, a core dump file is named
.IR core ,
but the
.I /proc/sys/kernel/core_pattern
file (since Linux 2.6 and 2.4.21)
can be set to define a template that is used to name core dump files.
The template can contain % specifiers which are substituted
by the following values when a core file is created:
.PP
.RS 4
.PD 0
.TP 4
%%
a single % character
.TP
%c
core file size soft resource limit of crashing process (since Linux 2.6.24)
.TP
%d
.\" Added in git commit 12a2b4b2241e318b4f6df31228e4272d2c2968a1
dump mode\(emsame as value returned by
.BR prctl (2)
.B PR_GET_DUMPABLE
(since Linux 3.7)
.TP
%e
executable filename (without path prefix)
.TP
%E
pathname of executable,
with slashes (\(aq/\(aq) replaced by exclamation marks (\(aq!\(aq)
(since Linux 3.0).
.TP
%g
(numeric) real GID of dumped process
.TP
%h
hostname (same as \fInodename\fP returned by \fBuname\fP(2))
.TP
%i
TID of thread that triggered core dump,
as seen in the PID namespace in which the thread resides
.\" commit b03023ecbdb76c1dec86b41ed80b123c22783220
(since Linux 3.18)
.TP
%I
TID of thread that triggered core dump, as seen in the initial PID namespace
.\" commit b03023ecbdb76c1dec86b41ed80b123c22783220
(since Linux 3.18)
.TP
%p
PID of dumped process,
as seen in the PID namespace in which the process resides
.TP
%P
.\" Added in git commit 65aafb1e7484b7434a0c1d4c593191ebe5776a2f
PID of dumped process, as seen in the initial PID namespace
(since Linux 3.12)
.TP
%s
number of signal causing dump
.TP
%t
time of dump, expressed as seconds since the
Epoch, 1970-01-01 00:00:00 +0000 (UTC)
.TP
%u
(numeric) real UID of dumped process
.PD
.RE
.PP
A single % at the end of the template is dropped from the
core filename, as is the combination of a % followed by any
character other than those listed above.
All other characters in the template become a literal
part of the core filename.
The template may include \(aq/\(aq characters, which are interpreted
as delimiters for directory names.
The maximum size of the resulting core filename is 128 bytes (64 bytes
in kernels before 2.6.19).
The default value in this file is "core".
For backward compatibility, if
.I /proc/sys/kernel/core_pattern
does not include
.I %p
and
.I /proc/sys/kernel/core_uses_pid
(see below)
is nonzero, then .PID will be appended to the core filename.
.PP
Paths are interpreted according to the settings that are active for the
crashing process.
That means the crashing process's mount namespace (see
.BR mount_namespaces (7)),
its current working directory (found via
.BR getcwd (2)),
and its root directory (see
.BR chroot (2)).
.PP
Since version 2.4, Linux has also provided
a more primitive method of controlling
the name of the core dump file.
If the
.I /proc/sys/kernel/core_uses_pid
file contains the value 0, then a core dump file is simply named
.IR core .
If this file contains a nonzero value, then the core dump file includes
the process ID in a name of the form
.IR core.PID .
.PP
Since Linux 3.6,
.\" 9520628e8ceb69fa9a4aee6b57f22675d9e1b709
if
.I /proc/sys/fs/suid_dumpable
is set to 2 ("suidsafe"), the pattern must be either an absolute pathname
(starting with a leading \(aq/\(aq character) or a pipe, as defined below.
.SS Piping core dumps to a program
Since kernel 2.6.19, Linux supports an alternate syntax for the
.I /proc/sys/kernel/core_pattern
file.
If the first character of this file is a pipe symbol (\fB|\fP),
then the remainder of the line is interpreted as the command-line for
a user-space program (or script) that is to be executed.
.PP
Since kernel 5.3.0,
.\" commit 315c69261dd3fa12dbc830d4fa00d1fad98d3b03
the pipe template is split on spaces into an argument list
.I before
the template parameters are expanded.
In earlier kernels, the template parameters are expanded first and
the resulting string is split on spaces into an argument list.
This means that in earlier kernels executable names added by the
.I %e
and
.I %E
template parameters could get split into multiple arguments.
So the core dump handler needs to put the executable names as the last
argument and ensure it joins all parts of the executable name using spaces.
Executable names with multiple spaces in them are not correctly represented
in earlier kernels,
meaning that the core dump handler needs to use mechanisms to find
the executable name.
.PP
Instead of being written to a disk file, the core dump is given as
standard input to the program.
Note the following points:
.IP * 3
The program must be specified using an absolute pathname (or a
pathname relative to the root directory, \fI/\fP),
and must immediately follow the '|' character.
.IP *
The command-line arguments can include any of
the % specifiers listed above.
For example, to pass the PID of the process that is being dumped, specify
.I %p
in an argument.
.IP *
The process created to run the program runs as user and group
.IR root .
.IP *
Running as
.I root
does not confer any exceptional security bypasses.
Namely, LSMs (e.g., SELinux) are still active and may prevent the handler
from accessing details about the crashed process via
.IR /proc/[pid] .
.IP *
The program pathname is interpreted with respect to the initial mount namespace
as it is always executed there.
It is not affected by the settings
(e.g., root directory, mount namespace, current working directory)
of the crashing process.
.IP *
The process runs in the initial namespaces
(PID, mount, user, and so on)
and not in the namespaces of the crashing process.
One can utilize specifiers such as
.I %P
to find the right
.I /proc/[pid]
directory and probe/enter the crashing process's namespaces if needed.
.IP *
The process starts with its current working directory
as the root directory.
If desired, it is possible change to the working directory of
the dumping process by employing the value provided by the
.I %P
specifier to change to the location of the dumping process via
.IR /proc/[pid]/cwd .
.IP *
Command-line arguments can be supplied to the
program (since Linux 2.6.24),
delimited by white space (up to a total line length of 128 bytes).
.IP *
The
.B RLIMIT_CORE
limit is not enforced for core dumps that are piped to a program
via this mechanism.
.\"
.SS /proc/sys/kernel/core_pipe_limit
When collecting core dumps via a pipe to a user-space program,
it can be useful for the collecting program to gather data about
the crashing process from that process's
.IR /proc/[pid]
directory.
In order to do this safely,
the kernel must wait for the program collecting the core dump to exit,
so as not to remove the crashing process's
.IR /proc/[pid]
files prematurely.
This in turn creates the
possibility that a misbehaving collecting program can block
the reaping of a crashed process by simply never exiting.
.PP
Since Linux 2.6.32,
.\" commit a293980c2e261bd5b0d2a77340dd04f684caff58
the
.I /proc/sys/kernel/core_pipe_limit
can be used to defend against this possibility.
The value in this file defines how many concurrent crashing
processes may be piped to user-space programs in parallel.
If this value is exceeded, then those crashing processes above this value
are noted in the kernel log and their core dumps are skipped.
.PP
A value of 0 in this file is special.
It indicates that unlimited processes may be captured in parallel,
but that no waiting will take place (i.e., the collecting
program is not guaranteed access to
.IR /proc/<crashing-PID> ).
The default value for this file is 0.
.\"
.SS Controlling which mappings are written to the core dump
Since kernel 2.6.23, the Linux-specific
.IR /proc/[pid]/coredump_filter
file can be used to control which memory segments are written to the
core dump file in the event that a core dump is performed for the
process with the corresponding process ID.
.PP
The value in the file is a bit mask of memory mapping types (see
.BR mmap (2)).
If a bit is set in the mask, then memory mappings of the
corresponding type are dumped; otherwise they are not dumped.
The bits in this file have the following meanings:
.PP
.PD 0
.RS 4
.TP
bit 0
Dump anonymous private mappings.
.TP
bit 1
Dump anonymous shared mappings.
.TP
bit 2
Dump file-backed private mappings.
.TP
bit 3
Dump file-backed shared mappings.
.\" file-backed shared mappings of course also update the underlying
.\" mapped file.
.TP
bit 4 (since Linux 2.6.24)
Dump ELF headers.
.TP
bit 5 (since Linux 2.6.28)
Dump private huge pages.
.TP
bit 6 (since Linux 2.6.28)
Dump shared huge pages.
.TP
bit 7 (since Linux 4.4)
.\" commit ab27a8d04b32b6ee8c30c14c4afd1058e8addc82
Dump private DAX pages.
.TP
bit 8 (since Linux 4.4)
.\" commit ab27a8d04b32b6ee8c30c14c4afd1058e8addc82
Dump shared DAX pages.
.RE
.PD
.PP
By default, the following bits are set: 0, 1, 4 (if the
.B CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
kernel configuration option is enabled), and 5.
This default can be modified at boot time using the
.I coredump_filter
boot option.
.PP
The value of this file is displayed in hexadecimal.
(The default value is thus displayed as 33.)
.PP
Memory-mapped I/O pages such as frame buffer are never dumped, and
virtual DSO
.RB ( vdso (7))
pages are always dumped, regardless of the
.I coredump_filter
value.
.PP
A child process created via
.BR fork (2)
inherits its parent's
.I coredump_filter
value;
the
.I coredump_filter
value is preserved across an
.BR execve (2).
.PP
It can be useful to set
.I coredump_filter
in the parent shell before running a program, for example:
.PP
.in +4n
.EX
.RB "$" " echo 0x7 > /proc/self/coredump_filter"
.RB "$" " ./some_program"
.EE
.in
.PP
This file is provided only if the kernel was built with the
.B CONFIG_ELF_CORE
configuration option.
.\"
.SS Core dumps and systemd
On systems using the
.BR systemd (1)
.I init
framework, core dumps may be placed in a location determined by
.BR systemd (1).
To do this,
.BR systemd (1)
employs the
.I core_pattern
feature that allows piping core dumps to a program.
One can verify this by checking whether core dumps are being piped to the
.BR systemd\-coredump (8)
program:
.PP
.in +4n
.EX
$ \fBcat /proc/sys/kernel/core_pattern\fP
|/usr/lib/systemd/systemd\-coredump %P %u %g %s %t %c %e
.EE
.in
.PP
In this case, core dumps will be placed in the location configured for
.BR systemd\-coredump (8),
typically as
.BR lz4 (1)
compressed files in the directory
.IR /var/lib/systemd/coredump/ .
One can list the core dumps that have been recorded by
.BR systemd\-coredump (8)
using
.BR coredumpctl (1):
.PP
.in +2n
.EX
$ \fBcoredumpctl list | tail -5\fP
Wed 2017-10-11 22:25:30 CEST  2748 1000 1000 3 present  /usr/bin/sleep
Thu 2017-10-12 06:29:10 CEST  2716 1000 1000 3 present  /usr/bin/sleep
Thu 2017-10-12 06:30:50 CEST  2767 1000 1000 3 present  /usr/bin/sleep
Thu 2017-10-12 06:37:40 CEST  2918 1000 1000 3 present  /usr/bin/cat
Thu 2017-10-12 08:13:07 CEST  2955 1000 1000 3 present  /usr/bin/cat
.EE
.in
.PP
The information shown for each core dump includes the date and time
of the dump, the PID, UID, and GID  of the dumping process,
the signal number that caused the core dump,
and the pathname of the executable that was being run by the dumped process.
Various options to
.BR coredumpctl (1)
allow a specified coredump file to be pulled from the
.BR systemd (1)
location into a specified file.
For example, to extract the core dump for PID 2955 shown above to a file named
.IR core
in the current directory, one could use:
.PP
.in +4n
.EX
$ \fBcoredumpctl dump 2955 \-o core\fP
.EE
.in
.PP
For more extensive details, see the
.BR coredumpctl (1)
manual page.
.PP
To disable the
.BR systemd (1)
mechanism that archives core dumps, restoring to something more like
traditional Linux behavior, one can set an override for the
.BR systemd (1)
mechanism, using something like:
.PP
.in +2n
.EX
# echo "kernel.core_pattern=core.%p" > /etc/sysctl.d/50\-coredump.conf
# /lib/systemd/systemd\-sysctl
.EE
.in
.PP
.\"
.SH NOTES
The
.BR gdb (1)
.I gcore
command can be used to obtain a core dump of a running process.
.PP
In Linux versions up to and including 2.6.27,
.\" Changed with commit 6409324b385f3f63a03645b4422e3be67348d922
if a multithreaded process (or, more precisely, a process that
shares its memory with another process by being created with the
.B CLONE_VM
flag of
.BR clone (2))
dumps core, then the process ID is always appended to the core filename,
unless the process ID was already included elsewhere in the
filename via a
.I %p
specification in
.IR /proc/sys/kernel/core_pattern .
(This is primarily useful when employing the obsolete
LinuxThreads implementation,
where each thread of a process has a different PID.)
.\" Always including the PID in the name of the core file made
.\" sense for LinuxThreads, where each thread had a unique PID,
.\" but doesn't seem to serve any purpose with NPTL, where all the
.\" threads in a process share the same PID (as POSIX.1 requires).
.\" Probably the behavior is maintained so that applications using
.\" LinuxThreads continue appending the PID (the kernel has no easy
.\" way of telling which threading implementation the user-space
.\" application is using). -- mtk, April 2006
.SH EXAMPLE
The program below can be used to demonstrate the use of the
pipe syntax in the
.I /proc/sys/kernel/core_pattern
file.
The following shell session demonstrates the use of this program
(compiled to create an executable named
.IR core_pattern_pipe_test ):
.PP
.in +4n
.EX
.RB "$" " cc \-o core_pattern_pipe_test core_pattern_pipe_test.c"
.RB "$" " su"
Password:
.RB "#" " echo \(dq|$PWD/core_pattern_pipe_test %p \
UID=%u GID=%g sig=%s\(dq > \e"
.B "    /proc/sys/kernel/core_pattern"
.RB "#" " exit"
.RB "$" " sleep 100"
.BR "^\e" "                     # type control-backslash"
Quit (core dumped)
.RB "$" " cat core.info"
argc=5
argc[0]=</home/mtk/core_pattern_pipe_test>
argc[1]=<20575>
argc[2]=<UID=1000>
argc[3]=<GID=100>
argc[4]=<sig=3>
Total bytes in core dump: 282624
.EE
.in
.SS Program source
\&
.EX
/* core_pattern_pipe_test.c */

#define _GNU_SOURCE
#include <sys/stat.h>
#include <fcntl.h>
#include <limits.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>

#define BUF_SIZE 1024

int
main(int argc, char *argv[])
{
    int tot, j;
    ssize_t nread;
    char buf[BUF_SIZE];
    FILE *fp;
    char cwd[PATH_MAX];

    /* Change our current working directory to that of the
       crashing process */

    snprintf(cwd, PATH_MAX, "/proc/%s/cwd", argv[1]);
    chdir(cwd);

    /* Write output to file "core.info" in that directory */

    fp = fopen("core.info", "w+");
    if (fp == NULL)
        exit(EXIT_FAILURE);

    /* Display command\-line arguments given to core_pattern
       pipe program */

    fprintf(fp, "argc=%d\en", argc);
    for (j = 0; j < argc; j++)
        fprintf(fp, "argc[%d]=<%s>\en", j, argv[j]);

    /* Count bytes in standard input (the core dump) */

    tot = 0;
    while ((nread = read(STDIN_FILENO, buf, BUF_SIZE)) > 0)
        tot += nread;
    fprintf(fp, "Total bytes in core dump: %d\en", tot);

    fclose(fp);
    exit(EXIT_SUCCESS);
}
.EE
.SH SEE ALSO
.BR bash (1),
.BR coredumpctl (1),
.BR gdb (1),
.BR getrlimit (2),
.BR mmap (2),
.BR prctl (2),
.BR sigaction (2),
.BR elf (5),
.BR proc (5),
.BR pthreads (7),
.BR signal (7),
.BR systemd\-coredump (8)